TESTABLE SPECIFICATIONS
Data Input Range. The assessment tool must track 5-10 sperm flagella at one time using diluted samples. A minimum of 20 diluted samples will be tested to give a complete analysis. To test this specification, our algorithm will analyze a minimum of 20 sperm videos, with a successful trial defined as being able to track the head of the sperm and the same spot on its flagellum through all frames of a video.
End-User Interface. The assessment tool must be user-friendly, and will be evaluated by a minimum of 3 clinicians. For each clinician, a minimum 90% satisfaction rate must be achieved. To test this specification, each clinician will be trained to use our algorithm to characterize sperm motility in a video. After each use, the clinician’s experience with the algorithm will be assessed using a Likert scale (see Appendix A). A 90% satisfaction rate will correspond to a score of 4.5 averaged across each clinician tested.
Operating Speed. The execution time for one trial, as defined by the analysis of one video of 5-10 sperm, using the algorithm should take no longer than five minutes. A minimum of 20 sperm videos will be analyzed, and the operating speed will be averaged across trials to determine if the execution time is within five minutes. The minimum computer processor required to run our program is a 2.4 GHz Intel Core i5.
Performance Requirements. The assessment tool should characterize sperm sample motility with a minimum accuracy rate of 90%. To test this specification, each clinician will view a video and identify which sperm cells are hyperactivated by inspection. We will then use our algorithm to identify hyperactivated sperm in the video and compare the results with the clinician’s observations. For 20 videos with 5 sperm per video, a 90% success rate per clinician corresponds to our algorithm agreeing with the clinician’s observations 90 out of 100 times.
Spatial Resolution. The assessment tool must be able to detect changes in amplitude within +/- 6.21 pixels (the number of pixels that corresponds to 1 micron at a magnification of 40X). To test this specification, our algorithm will be applied to computer-generated simulations of 5-10 moving sperm. We will compare our algorithm’s position data with the position data from the simulations as set by the user.
NON-TESTABLE SPECIFICATIONS
Data Security. The assessment tool must have 100% security and data safety so that all data is only accessible by the user.
Memory. The computerized assessment tool should occupy a maximum of 5GB on the user’s computer to avoid memory storage that exceeds the standard for algorithm codes.
Production and Consumer Costs. The production cost should not exceed $50. The consumer cost should not exceed $500. We require that the user have access to MATLAB through an academic institution, which ensures the cost will be at most $200.
Data Input Range. The assessment tool must track 5-10 sperm flagella at one time using diluted samples. A minimum of 20 diluted samples will be tested to give a complete analysis. To test this specification, our algorithm will analyze a minimum of 20 sperm videos, with a successful trial defined as being able to track the head of the sperm and the same spot on its flagellum through all frames of a video.
End-User Interface. The assessment tool must be user-friendly, and will be evaluated by a minimum of 3 clinicians. For each clinician, a minimum 90% satisfaction rate must be achieved. To test this specification, each clinician will be trained to use our algorithm to characterize sperm motility in a video. After each use, the clinician’s experience with the algorithm will be assessed using a Likert scale (see Appendix A). A 90% satisfaction rate will correspond to a score of 4.5 averaged across each clinician tested.
Operating Speed. The execution time for one trial, as defined by the analysis of one video of 5-10 sperm, using the algorithm should take no longer than five minutes. A minimum of 20 sperm videos will be analyzed, and the operating speed will be averaged across trials to determine if the execution time is within five minutes. The minimum computer processor required to run our program is a 2.4 GHz Intel Core i5.
Performance Requirements. The assessment tool should characterize sperm sample motility with a minimum accuracy rate of 90%. To test this specification, each clinician will view a video and identify which sperm cells are hyperactivated by inspection. We will then use our algorithm to identify hyperactivated sperm in the video and compare the results with the clinician’s observations. For 20 videos with 5 sperm per video, a 90% success rate per clinician corresponds to our algorithm agreeing with the clinician’s observations 90 out of 100 times.
Spatial Resolution. The assessment tool must be able to detect changes in amplitude within +/- 6.21 pixels (the number of pixels that corresponds to 1 micron at a magnification of 40X). To test this specification, our algorithm will be applied to computer-generated simulations of 5-10 moving sperm. We will compare our algorithm’s position data with the position data from the simulations as set by the user.
NON-TESTABLE SPECIFICATIONS
Data Security. The assessment tool must have 100% security and data safety so that all data is only accessible by the user.
Memory. The computerized assessment tool should occupy a maximum of 5GB on the user’s computer to avoid memory storage that exceeds the standard for algorithm codes.
Production and Consumer Costs. The production cost should not exceed $50. The consumer cost should not exceed $500. We require that the user have access to MATLAB through an academic institution, which ensures the cost will be at most $200.